Vitamin D Supplements Fail to Move Key Longevity Biomarkers in Major Trial Analysis

Vitamin D has spent the past two decades as one of the most heavily promoted supplements on the planet, credited by enthusiasts with everything from stronger bones to sharper immunity and a slower aging clock. A new systematic review and meta-analysis of randomized controlled trials, published in BMC Endocrine Disorders, now delivers a sobering verdict on two of the most talked-about molecular players in the vitamin D story: irisin, the so-called exercise hormone, and sirtuin 1, or SIRT1, a protein intimately linked to cellular longevity. After pooling data from seven randomized controlled trials comprising nine treatment arms, researchers led by Helia Amouji of Maragheh University of Medical Sciences found that vitamin D supplementation produced no statistically significant change in circulating levels of either biomarker compared with placebo.

The numbers behind that headline conclusion deserve close inspection, because they illustrate both the promise and the pitfalls of modern evidence synthesis. For irisin, the pooled weighted mean difference was 1.84 nanograms per milliliter, with a 95 percent confidence interval stretching from minus 0.25 to 3.93 nanograms per milliliter and a p-value of 0.084. For SIRT1, the weighted mean difference was 1.57 nanograms per milliliter, with a confidence interval of minus 0.18 to 3.31 and a p-value of 0.078. In both cases the point estimates lean in a positive direction, hinting that vitamin D might nudge these biomarkers upward, but the confidence intervals cross zero, which means the observed shifts are statistically indistinguishable from random noise at the conventional threshold of significance.

What makes the null result particularly striking is the sheer biological plausibility of the hypothesis being tested. Irisin is a myokine released by skeletal muscle during exercise, originally identified as a cleaved fragment of the membrane protein FNDC5, and it has been implicated in the browning of white adipose tissue, improved insulin sensitivity, and broader metabolic remodeling. SIRT1, meanwhile, is an NAD-dependent deacetylase that senses cellular energy status and influences mitochondrial biogenesis, inflammation, and DNA repair, making it a central node in the biology of aging. Vitamin D, acting through its nuclear receptor, regulates the expression of hundreds of genes, and laboratory studies have suggested crosstalk between vitamin D signaling and both irisin production and sirtuin activity. If supplementing with the vitamin were going to meaningfully rewire human metabolism through these pathways, the trials included in this analysis should have caught at least a whisper of it.

The research team cast a wide net to make sure that whisper, if it existed, would be heard. They systematically searched Scopus, Web of Science, and PubMed from database inception through 15 June 2026, restricting inclusion to randomized controlled trials, the gold standard for causal inference in clinical research. Two reviewers independently performed data extraction and risk-of-bias assessment using the Cochrane RoB 2 tool, the current benchmark for evaluating the internal validity of randomized trials. To generate pooled effect sizes, the authors used a random-effects model, which assumes that the true effect may vary from study to study rather than assuming a single fixed value, an appropriate choice given the diversity of populations, doses, and durations across the included trials. Heterogeneity, the statistical signature of that between-study variation, was quantified with the I-squared statistic, and sensitivity and subgroup analyses were deployed to hunt down its sources.

And here is where the story gets genuinely interesting, because the heterogeneity was enormous. The I-squared value for irisin was 95.0 percent, and for SIRT1 it was 92.3 percent, figures that indicate the vast majority of the observed variance between studies reflects real differences among the trials rather than chance. In meta-analytic terms, an I-squared above 75 percent is considered substantial; values in the nineties are a loud alarm that the pooled estimate is averaging together studies that may be measuring subtly different phenomena in different people. This is precisely why the authors paired their headline result with a battery of subgroup analyses, and why they urge readers to treat the headline, not the subgroups, as the primary finding.

Those subgroup analyses did surface some provocative signals. Vitamin D supplementation was associated with a significant increase in circulating irisin among patients with type 2 diabetes mellitus, with a p-value of 0.029, and among studies conducted in Iran, with a p-value of 0.025. For SIRT1, significant increases emerged in trials lasting twelve weeks, at a p-value of 0.008, and in the subgroup receiving 7,000 international units per week, which also reached p equals 0.008. The authors are careful to flag that this last estimate rests on only a single effect size, making it exploratory at best. In other words, the tantalizing possibility that vitamin D might raise irisin in diabetics, or boost SIRT1 over a twelve-week course, cannot be ruled out, but the evidence base is too thin and too heterogeneous to support firm clinical claims.

This tension between aggregate null results and suggestive subgroup findings is a recurring theme in nutrition science, and it carries real lessons for how the public should interpret supplement research. Subgroup analyses divide already-small datasets into smaller slices, and each additional slice multiplies the risk of false positives, a statistical phenomenon related to the multiple comparisons problem. A significant p-value in one subgroup, drawn from a handful of trials, is a hypothesis generator, not a prescription. The authors of the new analysis say as much in their conclusion, noting that the subgroup findings should be interpreted cautiously because of the small number of studies, the substantial heterogeneity, and the limited evidence underlying several of the subgroup estimates. It is a refreshingly restrained stance in a field where single positive subgroups often get inflated into marketing slogans.

There is also a deeper mechanistic question lurking beneath the statistics. Circulating irisin levels are driven primarily by physical activity and muscle mass, and SIRT1 activity is tightly coupled to cellular NAD availability and energy stress, both of which respond powerfully to exercise, caloric balance, and sleep but only weakly, if at all, to passive nutrient supplementation. A hormone whose raison d’être is to signal that muscles are working may simply not respond to a pill that tells the body nothing about mechanical load. The null result, viewed through this lens, is less a disappointment than a confirmation of a growing consensus: molecules that mediate the benefits of exercise and energy restriction tend to be recruited by the actual behaviors, not by pharmacological shortcuts that mimic one input to a vastly complex system.

The practical implications are straightforward. For the general population without a diagnosed deficiency, this analysis offers no support for taking vitamin D with the expectation of raising irisin or SIRT1, and by extension no support for claims that supplementation will deliver the metabolic and anti-aging benefits those biomarkers are often invoked to represent. For people with type 2 diabetes, and for the specific dosing and duration patterns flagged in the subgroups, the door remains open for larger, better-designed trials to test whether a real effect exists in those contexts. Until such trials arrive, the burden of proof stays where the authors left it: with the hypothesis, not the supplement bottle.

What the study ultimately demonstrates is the value of rigorous, preregistered-style evidence synthesis in a supplement landscape crowded with overinterpretation. By combining only randomized controlled trials, applying the Cochrane RoB 2 tool uniformly, using random-effects models, quantifying heterogeneity honestly, and running leave-one-out sensitivity analyses to confirm that no single study was driving the pooled estimates, the research team has produced a result that is more trustworthy than any individual trial in the set. The finding that vitamin D does not significantly move circulating irisin or SIRT1 will not trend as easily as a miracle-nutrient headline, but it is exactly the kind of careful, unglamorous work that gradually separates what supplements actually do from what the wellness industry wishes they did. In the ongoing effort to understand how vitamin D interacts with the molecular machinery of metabolism and aging, this meta-analysis closes one appealing shortcut and, in doing so, points researchers back toward the harder, more promising questions of who, if anyone, might genuinely benefit.

Subject of Research: The effect of vitamin D supplementation on circulating irisin and SIRT1 biomarker levels

Article Title: No significant effect of vitamin D supplementation on circulating Irisin and SIRT1 levels: a systematic review and meta-analysis of randomized controlled trials

Article References: Amouji, H., Pam, P., Panahiyan, P., Safari, M., Hojati, A., & Moradi, S. (2026). No significant effect of vitamin D supplementation on circulating Irisin and SIRT1 levels: a systematic review and meta-analysis of randomized controlled trials. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02518-1

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02518-1

Keywords: vitamin D, irisin, SIRT1, meta-analysis, systematic review, randomized controlled trials, biomarkers, supplements, metabolism, type 2 diabetes, heterogeneity, endocrinology